In this paper, we propose a feature-based automatic modulation recognition scheme for nine digital modulation signals, such as 2FSK, 4FSK, MSK, BPSK, QPSK, 8PSK, 16QAM, 32QAM, and 64QAM. The method uses magnitude spectrum powered by integer and higher order cumulants as features. For feature classification, Gaussian mixture model (GMM)-based classifier is used. Simulation results are demonstrated to evaluate the proposed method.

Keywords: automatic modulation recognition; Gaussian mixture model; higher order cyclic cumulants